U.S. patent number 4,685,004 [Application Number 06/712,094] was granted by the patent office on 1987-08-04 for rotary head type pcm recording and reproduction method and system.
This patent grant is currently assigned to Hitachi, Ltd.. Invention is credited to Takao Arai, Masaharu Kobayashi, Takaharu Noguchi, Toshifumi Shibuya, Hiroaki Takahashi.
United States Patent |
4,685,004 |
Takahashi , et al. |
August 4, 1987 |
Rotary head type PCM recording and reproduction method and
system
Abstract
In a recording operation by a rotary head type PCM
recording/reproducing apparatus including at least two magnetic
heads, the odd-numbered sample data and the even-numbered sample
data are discriminatively extracted and stored. The stored data
corresponding to an amount of data outputted from the A/D converter
within a predetermined time is selectively read out and supplied to
the two heads to thereby record the data in such a manner that the
odd-numbered data and the even-numbered data are recorded
separately on a track basis or alternatively recorded separately on
a preceding half and a succeeding half of each track. The error
detecting and correcting code is prepared on the basis of the data
to be recorded in each track. In reproduction of the recorded data,
upon detection of a drop-out of the data from one of the two heads,
the drop-out data is corrected through interpolation procedure to
thereby reconstitute the original analog signal.
Inventors: |
Takahashi; Hiroaki (Tokyo,
JP), Kobayashi; Masaharu (Yokohama, JP),
Noguchi; Takaharu (Yokohama, JP), Shibuya;
Toshifumi (Yokohama, JP), Arai; Takao (Yokohama,
JP) |
Assignee: |
Hitachi, Ltd. (Tokyo,
JP)
|
Family
ID: |
12872076 |
Appl.
No.: |
06/712,094 |
Filed: |
March 15, 1985 |
Foreign Application Priority Data
|
|
|
|
|
Mar 19, 1984 [JP] |
|
|
59-50915 |
|
Current U.S.
Class: |
360/32;
G9B/20.049; G9B/20.014; 360/48 |
Current CPC
Class: |
G11B
20/1809 (20130101); G11B 20/10527 (20130101) |
Current International
Class: |
G11B
20/10 (20060101); G11B 20/18 (20060101); G11B
005/00 (); G11B 005/09 () |
Field of
Search: |
;360/24,48,32,47 ;369/59
;371/2,45,39,40 |
References Cited
[Referenced By]
U.S. Patent Documents
|
|
|
4224642 |
September 1980 |
Mawatori et al. |
4477844 |
October 1984 |
Nakano et al. |
|
Primary Examiner: Canney; Vincent P.
Attorney, Agent or Firm: Antonelli, Terry & Wands
Claims
We claim:
1. A signal recording/reproducing method in a rotary head type PCM
recording/reproducing apparatus which comprises a plurality of
recording and reproducing heads, a rotatable cylinder having said
heads mounted thereon, an analog-to-digital (A/D) converter for
sampling an audio signal and converting it to a digital signal, a
digital-to-analog (D/A) converter for converting a digital signal
to a analog signal, a recording signal processing circuit for
adding an error detecting and correcting code to the digital
signal, and a reproduced signal processing circuit for performing
decoding and other processings of the reproduced signal, wherein
the output signal from said recording signal processing circuit is
recorded on a recording tape or the signal recorded on said tape is
reproduced, said method comprising, in recording mode, the steps
of:
(a) storing the odd-numbered sample data and the even-numbered
sample data discriminatively from the sampled digital signal
outputted from said A/D converter;
(b) reading selectively units of said stored data each
corresponding to a completed data set of a predetermined time
duration outputted from said A/D converter, said data units being
supplied to said plural heads to be recorded on said tape in such a
manner that the odd-numbered data and the even-numbered data are
recorded alternately on a plurality of tracks of said tape or
alternatively recorded on a preceding half and a succeeding half,
respectively, of each of said tracks;
(c) creating an error detecting and correcting code for the data to
be recorded on the track allotted thereto on the basis of the
stored data to be arrayed on said allotted track before the
recording of said read data on the tape, and recording said error
detecting and correcting code on the tape together with the
associated stored data; and
comprising, in reproducing mode, the steps of:
(d) storing the data signals reproduced from said recording tape
through said plural heads as the odd-numbered data and the
even-numbered data;
(e) reading selectively said reproduced and stored data for
converting to the analog signal to thereby reconstitute the
original audio signal; and
(f) checking the reproduced signal supplied from said plural heads
as to whether any of said plural tracks drops out, and upon
detection of the drop-out of the track, creating a correcting
signal for the lost data to be otherwise present on said drop-out
track through data interpolation based on the output data
reproduced from the other tracks.
2. A signal recording/reproducing method according to claim 1,
wherein said data unit is recorded on two adjacent tracks, said
rotatable cylinder having two diametrically disposed magnetic heads
mounted thereon.
3. A signal recording/reproducing method according to claim 1,
wherein data of the left channel signal or the right channel signal
of said audio signal which is a stereophonic signal is recorded
during a preceding half-period of each scan of said two heads,
while data of the right channel signal or the left channel signal
is recorded during a succeeding half-period, the relation between
said preceding half-period and said succeeding half-period being
reversed for the data recording during a succeeding scan.
4. A rotary head type PCM signal recording method in which an audio
signal is sampled to be converted into a digital signal, and an
error detecting and correcting code and attendant signals are added
to said digital signal, the resultant data being interleaved to be
recorded on a recording medium through a plurality of heads,
comprising the steps of:
(a) storing the odd-numbered sample data and the even-numbered
sample data discriminatively from the sampled digital signal;
(b) reading selectively units of said stored data each
corresponding to a completed data set of a predetermined time
duration of said digital signal, said data units being supplied to
said plural heads to be recorded on said tape in such a manner that
the odd-numbered data and the even-numbered data are recorded
alternately on a plurality of consecutive tracks of said tape;
(c) creating an error detecting and correcting code for the data to
be recorded on the track allotted thereto on the basis of the
stored data to be arrayed on said allotted track before the
recording of said read data on the tape, and recording said error
detecting and correcting code on the corresponding track of the
tape together with the associated stored data.
5. In a rotary head type PCM recorder which includes a pair of
magnetic heads, a rotatable cylinder having said magnetic heads
mounted thereon so as to be diametrically opposed to each other, an
analog-to-digital (A/D) converter for sampling an analog signal and
converting it to a digital signal, and a signal processing circuit
for adding to said digital signal an error detecting and correcting
code and predetermined signals, wherein the output of said signal
processing circuit is recorded on a magnetic recording medium and
the recorded data is reproduced from said magnetic recording
medium, a recording/reproducing method of recording or reproducing
said data comprising recording the odd-numbered sample data through
one of said two magnet heads, and recording and reproducing the
even-numbered sample data through the other magnetic head, said
odd-number sample data and said even-number sample data being
recorded alternately on a plurality of consecutive tracks of said
tape.
6. A recording/reproducing method according to claim 5, wherein the
digital data outputted from said A/D converter for a predetermined
period has a capacity capable of being recorded through two scans
of the magnetic heads.
7. A signal recording/reproducing method according to claim 6,
wherein said data error detecting and correcting code has a data
capacity which can be recorded through a single scan of said
magnetic heads.
8. A signal recording/reproducing method according to claim 5,
wherein the left channel or right channel data is recorded during
an earlier half-period of each scan of the two magnetic heads with
the right channel or left channel data being recorded during a
later half-period of said scan, while in a succeeding scan, the
relation between said earlier half-period and said later
half-period is reversed in such a manner that data of the right or
left channel is recorded during the earlier half period with data
of left or right channel being recorded during the later
half-period.
9. A rotary head type PCM recording/reproducing apparatus which
comprises a plurality of recording and reproducing heads, a
rotatable cylinder having said heads mounted thereon, an
analog-to-digital (A/D) converter for sampling an audio signal and
converting it to a digital signal, a recording signal processing
circuit for adding an error detecting and correcting code to the
digital signal, a reproduced signal processing circuit for
processing the waveform of the digital signal produced from said
magnetic head and performing error detecting and correcting
processing, a D/A converter for converting the output of said
reproduced signal processing circuit to an analog signal, wherein
the output signal from said recording signal processing circuit is
recorded on a recording tape or the signal recorded on said tape is
reproduced, comprising:
(a) a first memory for storing the odd-numbered sample data and the
even-numbered sample data discriminatively from the sampled digital
signal outputted from said A/D converter;
(b) data read means for reading selectively units of said stored
data each corresponding to a completed data set of a predetermined
time duration outputted from said A/D converter and controlling the
reading operation of said first memory so that said data units are
supplied to said plural heads to be recorded on said tape in such a
manner that the odd-numbered data and the even-numbered data are
recorded alternately on a plurality of consecutive tracks of said
tape;
(c) error detecting and correcting code creating means coupled to
said recording signal processing circuit for creating an error
detecting and correcting code for the data to be recorded on the
track allotted thereto on the basis of the stored data to be
arrayed on said allotted track before the recording of said read
data on the tape, and recording said error detecting and correcting
code on the tape together with the associated stored data;
(d) a second memory for storing the data signals reproduced from
said recording tape through said plural heads as the odd-numbered
data and the even-numbered data;
(e) means for reading selectively said reproduced and stored data
from said second memory for restoring the digital signal produced
by said A/D converter, and supplying said digital signal to said
D/A converter; and
(f) drop-out compensating means provided in the reproduced signal
processing circuit for checking the reproduced signal supplied from
said plural heads as to whether any of said plural tracks drops
out, and upon detection of the drop-out of the track, creating a
correcting signal for the lost data to be otherwise present on said
drop-out through data interpolation based on the output data
reproduced from the other tracks, said correcting signal being
supplied to said D/A converter.
Description
BACKGROUND OF THE INVENTION
The present invention relates to rotary head type PCM
recording/reproduction. More particularly, the invention concerns a
rotary head type PCM recording/reproduction method and system
suited for concealment and correction of such data errors or
defects as produced by drop-outs of the recording medium, such as a
magnetic tape, and the jamming of the head gap.
In the rotary head type PCM recording/reproducing apparatus
(hereinafter also referred to simply as "R-DAT recorder"), the
amount of data which can be recorded on a recording medium per unit
area (i.e. recording density) is increased by a factor of about 10
(ten) when compared with other systems such as e.g. a stationary
head type multi-track PCM recorder. When viewed from the standpoint
of the requisite number of heads, the R-DAT recorder for which two
heads are generally sufficient is more advantageous in respect to
the manufacturing cost and reliability than the stationary head
type multi-track PCM recorder which usually requires 10 to 20
heads. However, in the case of the rotary-head type PCM recorder,
there may arise a problem in which no reproduced data is derived
through one of the heads due to a drop-out of the magnetic tape
serving as the recording medium, and a jamming of the head gap is
caused by deposition of dust and/or magnetic particles scaled off
from the tape, resulting in the amount of data to be reproduced
being then decreased to a half to degrade significantly the quality
of reproduction, such as the reproduced sound. In the worst case,
the reproduction itself may become impossible.
In this regard, an examination in more detail will be made
concerning the principle of recording/reproduction adopted in the
hitherto known rotary head type PCM recorder as well as the
problems inherent in such systems, with reference to FIGS. 1 to 6
of the accompanying drawings. FIG. 1 graphically illustrates
two-channel stereophonic signals and the sampling points thereof.
Referring to FIG. 1, the left-channel signal (represented by L) and
the right-channel signal (represented by R) are sampled
alternatively on a time-sequential basis in the order of L.sub.0
.fwdarw.R.sub.0 .fwdarw.L.sub.1 .fwdarw.R.sub.1 .fwdarw.L.sub.2
.fwdarw.R.sub.2 and so on. In other words, the two-channel signals
are multiplexed on a time division basis so to be transmitted as a
serial signal. FIG. 2 shows in a block diagram a general
arrangement of a conventional rotary head type PCM recorder. In the
figure, a reference numeral 1 denotes a left-channel signal input
terminal, 2 denotes a right-channel signal input terminal, 3
denotes a left-channel input amplifier, 4 denotes a right-channel
input amplifier, 5 denotes a multiplexer (MPX) for alternately
changing over the channels to be recorded, 6 denotes an
analog-to-digital (A/D) converter, 7 denotes a digital signal
processing circuit for the recording signal, 8 denotes a recording
output amplifier, 9 denotes a rotatable head cylinder, 10 denotes a
magnetic head for a first channel (referred to also as "ch-1
head"), 11 denotes a magnetic head for a second channel (referred
to also as "ch-2 head"), 12 denotes a magnetic tape, 13 denotes a
reproducing preamplifier, 14 denotes a signal processing circuit
for the reproduced signal, 15 denotes a digital-to-analog (D/A)
converter, 16 denotes a demultiplexer (DMPX) for alternately
changing over the reproduced channels, 17 denotes a left-channel
output amplifier, 18 denotes a right-channel output amplifier, 19
denotes an output terminal for the left-channel signal, and numeral
20 denotes an output terminal for the right-channel signal. In the
operation of this R-DAT recorder, the audio signals of the
left-channel and right-channel are adjusted in level by the
respective input amplifiers 3 and 4 to be subsequently subjected to
channel multiplexing by the MPX 5 in the sequential order of L, R,
L, R as shown in FIG. 1. The analog signal outputted from the MPX 5
is converted into a digital signal by the A/D converter 6 to be
supplied to the digital signal processing circuit 7 where two basic
processings are performed as mentioned below. First, the data
transmitted thereto on a time serial basis is stored in a memory
such as a RAM incorporated in the processing circuit, which data is
then read-out from the memory in a sequence differing from the
order in which the data has been written. This operation is
referred to as the interleaving operation which is a sort of data
dispersing processing for allowing the correct original signal to
be restored even when the data as reproduced suffers from drop-outs
or errors in the block. This interleaving processing is carried out
by a data delay/rearraying circuit which is referred to as the
interleaver. Secondly, the interleaved signal is so arrayed as to
constitute error detection and correction blocks each of which has
added thereto a code signal for error detection and correction such
as a Reed-Solomon code. The signal processed as mentioned above is
recorded on the magnetic tape 12 by the pair of rotary magnetic
heads 10 and 11 after having been amplified through the recording
amplifier 8. Upon reproduction, the signal picked up from the
magnetic tape 12 by means of the magnetic heads 10 and 11 is
amplified by the reproduction preamplifier 13 and undergoes, if
necessary, correction of waveform referred to as waveform
equalization, and is subsequently supplied to the digital signal
processing circuit 14 of the reproduction system, where the input
data is arrayed so as to constitute the data detection/correction
blocks, the data suffering errors is detected and corrected (or
concealed) and the sequence of the interleaved data is transformed
to the original time-series sequence with the aid of a
deinterleaver. This processing is referred to as the deinterleaving
operation. The reproduced signal which has undergone the
processings mentioned above is then converted to an analogue signal
by the D/A converter 15 and subsequently is divided into the
left-channel signal and right-channel signal by the demultiplexer
circuit 16, resulting in the original audio signals L.sub.out and
R.sub.out being produced at the respective output terminals 19 and
20. FIG. 3 shows a record pattern produced on a magnetic tape by
the rotary head type PCM recorder, as viewed in the direction
perpendicularly to the plane of the magnetic tape. In FIG. 3, a
reference numeral 21 denotes tracks recorded by the magnetic head
10 for the first channel ch-1, and 22 denotes tracks recorded by
the magnetic head 11 for the second channel ch-2. Because the heads
10 and 11 usually differ from each other is azimuth angle, it is
possible to dispense with the non-recording zone (conventionally
referred to as a guard band) normally required between the first
channel ch-1 and the second channel ch-2. FIG. 4 shows a magnetic
tape whose magnetic surface has a defect due to injury or
deposition of dust particles. More specifically, a numeral 23
denotes a defect the record produced due to deposition of dust or
the scaling off of magnetic particles from the tape. When the
playback head passes by the defective areas where a drop-out or
dust is present, the defects are reflected as errors in the
reproduced signal. However, these errors are dispersed among data
by virtue of the interleaving procedure described above, and can be
corrected or concealed with the aid of the error correction code,
so that the reproduced signal deinterleaved in succession to the
error correction can be reproduced free of the influence of the
drop-outs or defects. FIG. 5 of the accompanying drawings shows a
typical one of the hitherto known record formats in which data is
recorded in a block array through the interleaving procedure which
is effective for error correction, as elucidated above. More
specifically, FIG. 5 shows the record format of an amount of data
which can be recorded and/or reproduced through a single scan by a
single head. This capacity will be referred to as one field. As
will be seen in FIG. 5, one field is composed of 256 blocks each of
which is subdivided in the manner illustrated in FIG. 6. For
particulars, reference may be made to Kentaro Odaka's article "A
Rotary Head High Density Audio Tape Recorder" of "Technical Report
Of Japan Electronic Communication Association", Vol. 1. 82, No.
190, EA 82-46 51, Nov. 30, 1982. Referring to FIG. 6, a numeral 24
designates a synchronizing signal SYNC (consisting of eight bits),
25 designates any identify code and block address code (consisting
of 16 bits), 26 denotes a parity check code for data (consisting of
32 bits), 27 designates audio signal data (of 96 bits), and 28
designates a cyclic redundancy check (CRC) code (of 16 bits) for
the error check operation. The parity code 26 is constituted by two
parity words referred to as P parity and Q parity each of 16 bits.
Further, data 27 is composed of six words W.sub.0, W.sub.1,
W.sub.2, W.sub.3, W.sub.4 and W.sub.5 each of 16 bits. In this
manner, one block is constituted by 168 bits in total, and 256
blocks in turn constitute one field. Referring again to FIG. 5, a
format of interleaving will be described. It will be seen that any
given block includes those words which are dispersed with a
distance corresponding to 128 words. In other words, any two
adjacent data words in a given block are chronographically
distanced from each other for a time corresponding to 128 sampling
points. By applying the interleaving procedure of this type, even
such error which may take place over a succession of the blocks due
to the drop-out of the tape or deposition of dust is dispersed
among the blocks composed of the interleaved data words each
dispersed by 128 sampling points and can be corrected by making use
of the Reed-Solomon code and/or b-adjacent code as the error check
code. Additionally, it will be noted that the even-numbered data
such as L.sub.0, R.sub.0, L.sub.2, R.sub.2 and so forth (referred
to also as even data) and the odd-numbered data such as L.sub.1,
R.sub.1, L.sub.3, R.sub.3 and so forth (also referred to as odd
data) are grouped in a left half and a right half, respectively, of
one field, as viewed in FIG. 5. By virtue of the classification or
grouping of the words in this manner, even when an error
uncorrectable error takes place to such extent that a major part of
the data contained in the preceding (left) half or alternatively in
the succeeding (right) half of one field has been erased, the
original analogue signal can be nevertheless reproduced
approximately in the original form through a concealment procedure
based on the previous value holding method or mean value
interpolation method, so far as either the even data or odd data
can be reproduced. In this connection, the length or extent of data
over which the error concealment can be effected in this way is
referred to as the error concealment length. In the case of the
interleaving illustrated in FIG. 5, the error concealment length
corresponds to 128 blocks.
As will be appreciated from the above discussion, the error
correction or error concealment can be accomplished with a
relatively high efficiency for the error produced due to the
drop-out of the magnetic tape or deposition of dust thereon. The
signal format of this type which is so arranged that the processing
of all data contained in one format can be completed within a
single scan period of one magnetic head will be referred to as the
single-scan-completion type format or the single-field-completion
type format. It should be however pointed out that the hitherto
known format of the single-field-completion type suffers serious
disadvantages in case the magnetic gap of one of the paired
magnetic heads is jammed through deposition of dust or magnetic
particles scaled off from the tape. Although error due to the
jamming of the magnetic gap does not occur so frequently as burst
error due to the drop-out, self-recovery from the jammed state can
scarcely be expected. The jamming of the head occurring in the
playback operation immediately brings about significant disturbance
in the reproduced sound. Of course, the jamming occurring in the
recording operation will give rise to a drop-out on the tape. With
the hitherto known signal format, the signal can not completely be
picked up by one magnetic head on the assumption that the head is
rotated at 2000 r.p.m., with the result that the reproduced sound
is intermittently intercepted at a time interval of 15 ms.
Accordingly, the jamming of the head gap may provide a death-blow
to the operation of PCM recording/reproducing apparatus which has
been developed intrinsically with the objective of producing sound
reproduction of a high fidelity. There exists thus a great demand
for dealing successfully with this problem.
Concerning the present state of the R-DAT technology, reference may
be made to the Japanese periodical "Electronics" of Japan
Electronic Industries Association, Vol. 24, No. 10 (1984), p.p.
36-42, inter alia p.p. 40-42.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a recording method
for a rotary head type PCM recorder capable of realizing a data
entry (format) for audio data recorded on a magnetic tape which
format allows the audio data to be reproduced with an improved
fidelity even when a drop-out on the tape takes place.
Another object of the present invention is to provide a reproducing
method and apparatus for a rotary head type PCM recorder which
makes it possible to derive a reproduced analogue signal with an
adequate sound quality for practical applications even when no
signal is produced by one of the magnetic heads due to the jamming
of the magnetic gap thereof.
For accomplishing the objects stated above, it is proposed
according to an embodiment of the invention that among the data as
sampled, the odd-numbered data (or odd data) is recorded and
reproduced by one head provided for one channel, while the
even-number data (even data) is recorded and reproduced by the
other magnetic head for the other channel or alternatively the odd
data and the even data are recorded and reproduced in a preceding
half and a succeeding half, respectively, of each track scan for
two channels, wherein data is so formatted as to be completed in
two fields (i.e. in two tracks), while the code for error detection
and correction is so arrayed as to be completed on a single-field
(single-track) basis.
According to an embodiment of the invention described above, data
is completed through the scan by the two heads. Accordingly, even
when there should arise such a situation in which no output signal
is produced by one of the heads at all, it is possible to reproduce
at least approximately the original analogue signal on the basis of
the output signal from the other head by making use of the average
value interpolation or the like method. Besides, even when an error
which has occurred is concentrated to a certain region of the
recording medium due to a drop-out on the tape or damage of the
tape edge, it is possible to avoid degradation of the sound quality
being concentrated or confined to a particular channel.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a view graphically showing sampling points on an input
analog signal.
FIG. 2 is a schematic block diagram showing an arrangement of a
hitherto known rotary head type PCM recorder.
FIG. 3 is a view showing a pattern of data recorded on a magnetic
tape.
FIG. 4 is a view for illustrating, by way of example, a drop-out on
a magnetic tape.
FIG. 5 is a view illustrating a data recording format constituted
by one field.
FIG. 6 is a view showing a data array in one block contained in the
field.
FIG. 7 is a view illustrating a data recording pattern in which the
odd data and the even data are recorded on separate tracks,
respectively, on a magnetic tape according to an embodiment of the
present invention.
FIG. 8 is a view illustrating a data recording pattern in which
left channel data and right channel data are recorded at separate
areas, respectively, on a magnetic tape according to another
embodiment of the invention.
FIG. 9 is a view illustrating a format for recording data in two
fields.
FIG. 10 is a view illustrating division of data in two fields.
FIG. 11 is a view illustrating division of data in four fields.
FIG. 12 is a view illustrating division of data in two fields.
FIG. 13 is a view showing a modification of the data pattern shown
in FIG. 8.
FIG. 14A (consisting of FIGS. 14A-1 and 14A-2) and 14B (consisting
of FIGS. 14B-1 and 14B-2) are views illustrating in detail the
formats for recording data in two fields in accordance with the
data pattern shown in FIG. 13.
FIG. 15 shows in block diagram a general arrangement of a rotary
head type PCM recording/reproducing apparatus for carrying out the
recording/reproducing method according to the present
invention.
FIG. 16 is a view for illustrating a data storage structure of a
memory employed in the apparatus shown in FIG. 15.
FIGS. 17A and 17B show timing signal waveform diagrams for
illustrating recording and reproducing operations,
respectively.
FIG. 18 is a schematic diagram showing an example of a switching
circuit for controlling read/write operations of the memory shown
in FIG. 16.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Now, the invention will be described in detail in conjunction with
the preferred embodiments thereof.
FIG. 7 is a view for illustrating a recording/reproducing method
for a rotary head PCM recorder according to an exemplary embodiment
of the invention. In the figure, a reference numeral 24 designates
skew tracks on which only the sampled data i.e. data words
(referred to as odd data) that correspond to the odd-numbered
sampling points are recorded, and 25 designates skew tracks on
which only the sampled data (referred to as even data)
corresponding to the even-numbered sampling points are recorded. As
will be seen in FIG. 7, the odd data is recorded on the tracks
allocated to the magnetic head for a first channel ch-1 with the
even data being recorded on the tracks allocated to the magnetic
head for a second channel ch-2. By dividing or classifying the data
in this manner, it becomes possible to reproduce the original
analog signal without causing any appreciable discomfort to the
ears of listeners, even when one of the heads does not at all
produce an output signal as a result of jamming of the magnetic gap
thereof. This is because either the even data signal or the odd
data signal among the sampled and recorded data signals can be
reproduced from the tracks provided separately for the magnetic
heads, whereby the lost data, e.g. the odd data can be correctly
reconstituted through average value interpolation based on the even
data and vice versa. In this connection, the detection of a
drop-out due to the jamming of the magnetic gap or other causes can
be performed by using the error check code known per se.
In the case of the recording format shown in FIG. 8 in which the
odd data and the even data are recorded separately and alternately
on the tracks, it is further proposed according to the invention
that the left channel data L is recorded in a preceding half of a
given single scan (single field) of each head while the right
channel data R is recorded in the succeeding half of the scan. On
the other hand, during the next single scan, the right channel data
R is recorded in the preceding half, which is followed by the
succeeding half in which the left channel data L is recorded. With
this structure of the format, it is possible to reconstitute the
lost sampled data through the mean value interpolation procedure,
even when error should be concentrated on the data recorded during
the preceding half or during the succeeding half of one scan of
each head due to deposition of dust on the magnetic tape surface or
damage to the tape such as breakage of the tape edge or twisting of
the tape.
The recording/reproducing method described above will be elucidated
more in detail in connection with the data recording format. FIG. 9
shows an example of the data recording format which can be adopted
in carrying out the recording/reproducing method according to the
invention. In FIG. 9, a reference letter 29A designates a single
field containing only the odd data, and 29B designates another
single field containing only the even data. More specifically, the
data in the one field 29A is recorded on one of the tracks 24 shown
in FIG. 7 while the data in the other single field 29B is recorded
on one of the tracks 25. In the case of the format shown in FIG. 9,
one data block is constituted by units referred to as symbols each
of several bits. In consideration of the requirement imposed from
the standpoint of hardware, it is common that one symbol consists
of 4 to 8 bits. In the case of the illustrated embodiment, it is
assumed that one symbol consists of eight bits. Accordingly, when
an analog signal is sampled, for example, at 16 quantizing bits,
the data of the analog signal at a given sampling point (usually
referred to as one word data) can be represented by two symbols.
Now referring to FIG. 9, one block is constituted by a
synchronizing signal SYNC (of one symbol=1 byte), an identify code
signal IDENTIFY (of three symbols) for address and control data,
data signal or second decoding parity signal C2-PARITY making use
of the Reed-Solomon code (of 30 symbols), and an error detection
signal C1-PARITY (of 2 symbols). Thus, one block consists of 36
symbols or 288 bits in total. On the other hand, one field is
composed of 128 blocks of which the first 48 blocks and the last 48
blocks (hence 96 blocks in total) are allocated for the recording
of data signal, while 32 intermediate blocks are allotted for
recording the parity signal C2-PARITY. In the case of the
illustrated embodiment, the odd data the left channel L-ch (or
right channel) is arrayed in the first 48 blocks of one field 29A,
while the odd data of the right channel R-ch (or left channel) is
located in the last 48 blocks. In the succeeding adjacent field
29B, the even data of the right channel R-ch (or left channel L-ch)
is recorded in the first 48 blocks while the even data of left
channel L-ch (or right channel R-ch) is located in the last 48
blocks. By virtue of the data array described above, it is possible
to derive a reproduced signal approximating the original analog
signal through the average value interpolation procedure, even when
either one of the head ch-1 or ch-2 produces no output, because at
least one of the even or odd data recorded on the mutually
different tracks can be reproduced without fail. However, with
regard to the block structure of the error detection and correction
code used in connection with the recording of the data
corresponding to an analog signal of a given length over or across
two fields or tracks, the error detection and correction code
should desirably be completed for each track, i.e. within one
field. In other words, even if the data signal derived by sampling
the original signal is present across two field, it is required
that the error detection/correction signal (parity signal) which is
created through arithmetic processing of the data signal should be
completed within one field. This is because the data of one field
has to be correctly reconstituted within the one field, even when
either head ch-1 or ch-2 should fail to produce an output, so long
as the other head produces the data signal as well as the parity
signal. The data format shown in FIG. 9 is prepared to meet the
above requirement, wherein the Reed-Solomon code is made use of as
the error detection/correction code. To this end, the data word of
24 symbols is encoded by using (32, 34) Reed-Solomon codes.
Subsequently, the symbols constituting the C2-parity code block are
interleaved, and the data word or C2-parity word of 30 symbols
constituting one block is encoded by using (32, 20) Reed-Solomon
code. Consequently, the pair of parity signals C1-PARITY and
C2-PARITY are formed in the fields 29A and 29B independently, which
means that the code structure is completed in each field.
Next, another embodiment of the invention will be described in
which two signals sampled at mutually different sampling
frequencies are recorded. In the rotary head type PCM recorder, the
format has to be such that the digital dubbing of data derived
through the sampling at the frequencies of 48 KHz and 32 KHz (as
used in direct satellite broadcasting or DBS) and 44.1 KHz (used in
the compact disk system) can be effected. The recording of data
sampled at the frequencies of 48 KHz and 32 KHz is carried out by
varying correspondingly the transportation speed of the tape. On
the other hand, for the recording of data resulting from the
sampling at the frequency of 44.1 KHz, a format for recording data
sampled at 48 KHz is used in common. However, in the case of the
data obtained through the sampling at the frequency of 44.1 KHz and
containing a smaller number of data words than the data sampled at
48 KHz, there is formed a blank region from which no data is read
out during the playback operation. In FIG. 9, the amount of data
recorded in two fields is equal to 2800 words (each containing of
16 bits). Assuming that this capacity corresponds to the amount of
data derived through sampling at 48 KHz, the capacity required for
recording the data sampled at the frequency of 44.1 KHz is
proportionately decreased to 2646 words.
FIG. 10 shows the allocation of data resulting from the sampling at
48 KHz to the field assigned to the heads ch-1 and ch-2,
respectively. More specifically, 2880 words in total inclusive of
1440 words for the left and right channel data, respectively, are
allocated to two fields. In the field assigned to the head ch-1,
there are recorded the odd data of 720 words out of 1440 words
constituting the left channel data and the odd data of 720 words
out of the 1440 words constituting the right channel data, while
the remaining even data of the right and left channels are recorded
in the field assigned to the magnetic head ch-2.
On the other hand, for recording data derived through the sampling
at the frequency of 44.1 KHz, there are required 2646 words in
total inclusive of the 1323 words of the left channel data and 1323
words of the right channel data, wherein 661 words of the
left-channel odd data and 661 words of the right-channel odd data
are recorded in the field assigned to the magnetic head ch-1. There
are recorded in the succeeding field assigned to the head ch-2 662
words of the left-channel even data and 662 words of the
right-channel even data. Thus, data is completed in the two fields
mentioned above. However, there remains an unbalance between the
odd data and the even data. Accordingly, 662 words of the
left-channel odd data and 662 words of the right-channel odd data
are recorded in the next field assigned to the head ch-1, while 661
words of the right-channel even data and 661 words of the
left-channel even data are recorded in a further succeeding field
for the head ch-2. FIG. 11 shows the format thus prepared, wherein
illustrating the locations at which 2646 words of data words 0 to
2645 are recorded. In the case of the instant embodiment, the field
assigned to the head ch-1 is assumed to record only the odd data
with the field assigned to the head ch-2 recording only the even
data. It will however be understood that the change in the number
of data words derived through the sampling at the frequency of 44.1
KHz can be dealt with by changing over the number of words to be
processed between 661 and 662.
FIG. 12 shows a recording format prepared for the data sampled at
the frequency of 44.1 KHz according to another embodiment of the
invention. In the case of this format, it is assumed that the head
ch-1 is used for recording the odd data with the head ch-2 being
employed for the recording the even data, wherein 661 words of the
left-channel data are recorded in the field for the head ch-1 with
662 words of the same channel being recorded in the field ch-2,
while 662 words of the right-channel data are recorded in the field
assigned to the head ch-1 with the 661 words of the same channel
being recorded in the field for the head ch-2. With the structure
of the format shown in FIG. 12, 1323 words (sampled data) for the
left and right channels, respectively, are obtained as the complete
data through two scans by the heads ch-1 and ch-2.
It should be appreciated that the invention is never restricted
exactly to the formats and the numerical values described above,
but it can be implemented in other format structures in combination
with other numerical values or dimensions, to similar effect.
FIG. 13 shows a modification of the recording pattern shown in FIG.
8 which is rendered more immune to error due to a drop-out on the
tape. More specifically, in the format shown in FIG. 8 in which the
left channel data and the right channel data are alternately
recorded in the preceding (earlier) and succeeding (later) halves,
respectively, of the track as observed in the head scan direction
of the track row, the format shown in FIG. 13 is so modified that
the odd data and the even data are recorded, being allocated to the
preceding and succeeding halves, respectively, of each scan of the
head. Consequently, in the playback operation, the complete data
can be retrieved from the paired adjacent tracks. For example, the
tracks (a) and (b) are scanned to read-out sequentially the
right-channel odd data, the left-channel even data, the
left-channel odd data and the right-channel even data, which are
subsequently stored at the memory locations described below. Then,
by controlling the sequence in which this data is read-out from the
memory, the audio data arrays for the left channel and the right
channel are reconstituted, to thereby reproduce the original
stereophonic audio signals. Since in the rotary head type PCM
recorder, data is recorded with different azimuth angles on any
adjacent tracks, as is well known in the art, the paired track can
be easily identified. Of course, it is possible to allocate the
preceding half of a single head scan to the recording of the even
data with the succeeding half being allocated to the recording of
the odd data.
FIGS. 14A and 14B are views illustrating an example of the data
recording format which can be employed in carrying out the
recording/reproducing method in accordance with the modified
recording pattern shown in FIG. 13. More particularly, there are
shown in FIGS. 14A and 14B the contents of the paired fields 29A'
and 29B' having recorded therein the odd data and the even data,
respectively. The data of the first one 29A' of the paired fields
are recorded on the track (a) shown in FIG. 13, while the data of
the second field 29B' are recorded on the track (b). With the
format illustrated in FIGS. 14A and 14B, the data blocks described
hereinbefore in conjunction with FIG. 9 are formed. For the signals
sampled from the right-channel and left-channel audio signals, L
and R in FIGS. 14A and 14B represent the left- and right-channel
signals, the affixed numerals represent the sampling sequence, and
the affixed letters u and l represent the more significant and less
significant bits or symbols. Letters P and Q represent the
Reed-Solomon codes C1-PARITY and C2-PARITY, respectively. The
sampled time-serial signals L and R are in the state of having
undergone processing for forming the error detection and correction
signal as well as the interleaving processing described
hereinbefore in conjunction with the format shown in FIG. 9.
It should here be mentioned that the parity symbols are defined so
as to satisfy the following equations:
where H.sub.Q and Hp represent parity matrixes given by ##EQU1##
The calculation is defined on GF(2.sup.8) by the following
polynomial.
and a primitive element .alpha. in GF(2.sup.8) is defined as
follows. ##EQU2##
FIG. 15 shows in a schmatic block diagram a general arrangement of
the rotary head type PCM recording/reproducing apparatus according
to an embodiment of the invention.
At first, description will be directed to the operation in the
recording mode.
A left-channel analog input signal and a right-channel analog
signal applied to input terminals 1 and 2 are selected by a
multiplexer or MPX 5 to be converted alternately to a digital
signal of 16 bits by an analog-to-digital (D/A) converter 6. The
digital signal of 16 bits is then divided into a signal of eight
more significant bits (MSB) and a signal of eight less significant
bits (LSB) through a 16-to-8 bit conversion circuit 31, which
signals are then alternately outputted onto a 8-bit bus line
30.
The signal consisting of eight MSBs and the signal consisting of
eight LSBs produced on the bus-line 30 are sequentially stored in a
memory 32 at predetermined locations or regions.
The signals stored in the memory are encoded by an error correcting
code encoder/decoder (ECU) 33 and subsequently read out in a
predetermined sequence to be modulated by a modulating circuit 34,
the output signal of which is amplified by a recording amplifier 8
and recorded on a magnetic tape 12 through magnetic heads 10 and
11.
Upon reproduction, the magnetic heads 10 and 11 are changed over to
the reproducing mode from the recording mode by a changing-over
switch (not shown). The reproduced signal outputted from the heads
is amplified by a preamplifier 13 and subjected to predetermined
waveform equalizing processing through a waveform equalizer 35.
After discrimination as to the bit value "1" or "0" has been
effected by a data strobe circuit 36, the signal is demodulated by
a demodulator circuit 37 to be produced onto the bus line 30.
The reproduced and demodulated signals on the bus line 30 are
sequentially stored in the memory 32 at predetermined locations or
regions.
The signals stored in the memory are decoded by the error
correcting code encoder/decoder 33 and subsequently read out in a
predetermined sequence. The signals consisting of eight MSBs and
eight LSBs as read out are converted into the original digital
signal of 16 bits by a 8-to-16 bit conversion circuit 39. After
having undergone the processing for concealing the uncorrectable
erroneous data through a decoding circuit 33, the 16-bit signal is
converted to an analog signal through a digital-to-analog (D/A)
converter 15 so as to be divided into the right-channel signal and
the left-channel signal. Alternatively, the output signal of the
D/A converter may be sampled and held by sample and hold circuits
17 and 18 alternately to be supplied to audio output terminals 19
and 20 as the left-channel and right-channel signals.
Next, the interleave processing which may be adopted in the instant
embodiment will be described by referring to a memory map shown in
FIG. 16.
The memory 33 is constituted by four storage regions M.sub.a,
M.sub.b, M.sub.c and M.sub.d, as shown in FIG. 16. These regions
are dealt with on the basis of the paired regions M.sub.a :M.sub.b
and M.sub.c :M.sub.d.
In the recording operation, the sampled signal data of left and
right (L-ch, R-ch) channels are so recorded that the odd data and
the even data are located at the respective allocated regions as
shown in FIG. 16.
For example, the memory is divided into four regions each of which
is then divided into three subregions.
At first, for fetching and outputting the data on a time-division
basis, data is divided and stored in two paired regions M.sub.a
:M.sub.b and M.sub.c :M.sub.d, where the memory regions M.sub.a and
M.sub.c on one hand and M.sub.b and M.sub.d on the other hand are
of the identical structure. There are stored in the regions M.sub.a
and M.sub.b the even data and the odd data of the left and right
channels, respectively, for a predetermined period. For example,
the even data of the left channel is recorded in the region M.sub.a
with the odd data of the same channel being recorded in the region
M.sub.b, while the odd data of the right channel is stored at the
region M.sub.a with the even data of the same channel being stored
at the region M.sub.b. Further, the even data of the left channel
is stored at the preceding (upper) half area of the region M.sub.a
with the odd data of the same channel being stored at the
succeeding (lower) half area of the region M.sub.b. The odd data of
the right channel is stored at the succeeding half area of the
region M.sub.a with the even data of the same channel being stored
at the preceding half area of the region M.sub.b. In each of the
regions, the parity code is stored at the respective intermediate
area.
The operation mentioned above is controlled by a memory control
circuit 38 on the basis of an address signal. Namely, in dependence
on the results of discrimination as to whether the data is of the
right or left channel and whether the data is odd or even, the
regions at which the data is to be recorded are controlled by using
the address signals.
Further, the encoding operation for error detection and correction
described above is executed for each of the regions M.sub.a and
M.sub.b. During a period in which the data stored in the regions
M.sub.a and M.sub.b is encoded by the ECU 33, the encoded data
stored at the regions M.sub.c and M.sub.b is recorded on the
magnetic tape 12. The encoding operation and the recording
operation are performed alternately.
In the reproducing operation, the reproduced signal data is fetched
by the memory 32 for reproduction at the regions M.sub.a, M.sub.b,
M.sub.c and M.sub.d in this order. The data thus fetched is decoded
by the ECU 33 in the sequence in which it has been fetched. The
decoded data is sequentially read out as left channel data and the
right channel data from the pair of regions M.sub.a and M.sub.b and
the pair of regions M.sub.c and M.sub.d, respectively.
In the illustrated embodiment, the memory for the recording
operation and the memory for the reproducing operation are
constituted by the same memory. It is however possible to provide
two memories one of which is constituted by the memory regions
M.sub.a and M.sub.b with the other memory being constituted by the
regions M.sub.c and M.sub.d.
Next, the operation of the recording system will be described by
referring to the timing waveform diagram shown in FIG. 17A. During
a period of time points t.sub.0 to t.sub.4, the memory regions
M.sub.a and M.sub.b are selected. By selecting the regions M.sub.a
and M.sub.b alternately, the odd and even data samples of the left
and right channels are fetched and loaded at predetermined regions
or areas of the memory (M.sub.a, M.sub.b). During the encoding
period from t.sub.4 to t.sub.5, data located in the memory region
M.sub.a is encoded. During the output period from t.sub.5 to
t.sub.6, the encoded data of the memory region M.sub.a is recorded
on the magnetic tape in the predetermined format during a recording
period (54). During a period from t.sub.7 to t.sub.8, the encoded
data of the memory region M.sub.b is recorded on the magnetic tape
in the predetermined format. During a period from t.sub.5 to
t.sub.8, the memory regions M.sub.c and M.sub.d are selected.
Subsequently, data is fetched, encoded and recorded on the magnetic
tape in a similar manner.
Next, the operation of the reproducing system will be described by
referring to the timing diagram illustrated in FIG. 17B. During a
data reproducing period (55) from t.sub.1 to t.sub.5, the
reproduced data is loaded in the memory region M.sub.a as in the
case of the data fetching period (56). During a decoding period
(57) from t.sub.2 to t.sub.3, the data is decoded. During a data
fetching period (56) from t.sub.3 to t.sub.4, the reproduced data
is loaded in the memory region M.sub.b. During the decoding period
(57) from t.sub.4 to t.sub.5, the data is decoded. During an output
period (58) from t.sub.5 to t.sub.9, the decoded data of the memory
regions M.sub.a and M.sub.b are sequentially outputted on the
time-serial basis.
Further, during the data reproducing periods (55) from t.sub.5 to
t.sub.6 and from t.sub.7 to t.sub.8, the reproduced data is stored,
decoded and outputted sequentially on the time-series basis.
FIG. 18 is a view showing a switch circuit for illustrating a
method of controlling the read/write operations to and from the
memory 32, by way of example. In operation, a switch SW-1 is closed
to the memories M.sub.a and M.sub.b while a switch SW-10 is closed
to the memories M.sub.c and M.sub.d during a period from t.sub.0 to
t.sub.5. Subsequently, during a period from t.sub.0 to t.sub.1,
switches SW-2 and SW-3 select the memories M.sub.a or M.sub.b in
dependence on the result of identification of the data supplied to
the memory input terminal M.sub.IN as to the left or right channel
and odd or even data. During a period from t.sub.0 to t.sub.1, a
switch SW-8 is closed to decode the data of the memory region
M.sub.c, the decoded data being outputted during a succeeding
period from t.sub.1 to t.sub.2. Similarly during a period from
t.sub.2 to t.sub.3, a switch SW-9 is closed to decode the data
located at the memory region M.sub.d, the decoded data being
outputted during a succeeding period from t.sub.3 to t.sub.4. The
data outputted through the switches SW-8 and SW-9 is transferred to
the memory output terminal M.sub.OUT by way of the switch SW-10
closed to the memory (M.sub.a, M.sub.b) during a period from
t.sub.0 to t.sub.4.
In a similar manner, data input and output are repeated alternately
between the memories M.sub.a :M.sub.b and M.sub.c :M.sub.d.
Next, the operation for reproduction will be described.
The reproduced signal is inputted to the memory input terminal
M.sub.IN, while the memory output signal is outputted from the
memory output terminal M.sub.OUT. At first, during a period from
t.sub.1 to t.sub.5, the switch SW-1 is closed to the memories
M.sub.a and M.sub.b. During a period from t.sub.1 to t.sub.2, the
switch SW-2 is closed to the memory M.sub.a, whereby the reproduced
signal is stored in the memory M.sub.a. During a period from
t.sub.2 to t.sub.3, the signal placed in the memory M.sub.a is
decoded. During a period from t.sub.3 to t.sub.4, the switch SW-3
is closed to the memory M.sub.b, whereby the reproduced signal is
stored in the memory M.sub.b. During a period from t.sub.4 to
t.sub.5, the signal located in the memory M.sub.b is decoded.
Subsequently, during a period from t.sub.5 to t.sub.9, the switch
SW-10 is closed to the memories M.sub.a and M.sub.b, while the
switches SW-6 and SW-7 are alternately closed, resulting in the
signals stored in the memories M.sub.a and M.sub.b being outputted
sequentially on a time-series basis.
Thereafter, the data reading/writing operation is repeated
alternately between the memories M.sub.a :M.sub.b and the memories
M.sub.c :M.sub.d in a similar manner.
In the foregoing, the operation of the memories has been described
on the assumption that the switches SW-1 to SW-10 are used. It will
however be readily understood by those skilled in the art that the
processing based on a similar operation can be performed by using a
single or plural memories under the address control on a
time-division basis.
* * * * *